Materials Science-Poland, Vol. 23, No.

3, 2005

Preparation of CuFe2O4/SiO2 nanocomposite by the sol-gel method

J. PLOCEK1, 2*, A. HUTLOV1, D. NIANSK1, 2, J. BURK1, J.-L. REHSPRINGER3, Z. MIKA2
1 2

Institute of Inorganic Chemistry, ASCR, 25068 e u Prahy, Czech Republic

Department of Inorganic Chemistry, Faculty of Science, Charles University of Prague, Albertov 2030, 128 40 Prague 2, Czech Republic

I.P.C.M.S., Groupe des Matriaux Inorganiques, 23 Rue du Loess, F-67037 Strasbourg Cedex, France

This work aims at characterizing the phase relations in the CuFe2O4/SiO2 system. Samples were prepared by the sol-gel method. Final heat treatment of the samples was carried out at temperatures in the range of 8001100 C. Final products were characterized by HR TEM, X-ray diffraction, magnetic measurements, and Mssbauer spectroscopy. HR TEM revealed nanocrystals with sizes of 7130 nm, depending on the heat treatment temperature. The spinel structure of CuFe2O4 in the amorphous silica matrix proved to be stable up to 1100 C without decomposition to copper silicate and iron (III) oxide. At the same time, the amorphous silica matrix recrystallized to cristobalite at 1100 C. Key words: sol-gel; spinel ferrite; silica matrix; nanocomposite; Mssbauer spectroscopy; transmission electron microscopy; magnetic measurements

1. Introduction
Nanocomposites have been the subject of many studies in recent years due to the new properties they are expected to exhibit [1]. One of the interesting groups consists of metal oxide compounds in a silica matrix. These materials can have interesting magnetic and magnetooptical properties [2]. One of the ways to prepare nanocomposites with the required properties is the sol-gel method. The advantages of this method are better homogeneity of materials and lower temperatures of treatment. In the case of nanocomposites in a silica matrix, samples with an arbitrary cation to silica ratio can be prepared and the particle size can be controlled by the parameters of heat treatment. _________
*

Corresponding author, e-mail: plocek@ iic.cas.cz

698

J. PLOCEK et al.

This method was, for example, successful in the preparation of a metastable phase in the Fe2O3/SiO2 system [35]. Studies have revealed that the metastable -Fe2O3 could be present in this nanocomposite up to 800 C and that the intermediate phase -Fe2O3 also appears [6]. In the case of ternary oxides with a spinel structure, the situation is more complicated due to the possible formation of silicate. The MFe2O4/SiO2 system is metastable from the theoretical point of view (M2+ silicates are formed). Previous studies have revealed, however, that the spinel phase was formed in the silica matrix and that the formation of MFe2O4 nanoparticles strongly depends on the type of M2+. Spinel nanocrystals in the silica matrix are stable up to 1100 C in the case of M = Co, Ni, and Zn [7, 8] and mainly cadmium silicate and iron(III) oxide were found in the final heat treated product in the case of Cd [8]. This work presents the preparation of another MFe2O4/SiO2 system (here M = Cu) and the characterization of the phase relations in this system. It also concerns the ferrite/silica-nanocomposite system prepared by the sol-gel method, which has not yet been studied. This work aims to show the suitability of this method for preparing copper ferrite nanoparticles in the silica matrix. CuFe2O4/SiO2 nanocomposites were prepared by the sol-gel method and were heat treated in the temperature range of 8001100 C. The final products were studied by X-ray diffraction, HR TEM, magnetic measurements, and Mssbauer spectroscopy.

2. Experimental
Sample preparation. Samples were prepared using the conventional sol-gel method. Copper and iron nitrates were used as spinel precursors. TEOS, HNO3 as an acid catalyst, formamide as a modifier, and methanol as a solvent were employed for silica matrix preparation. Fe(NO3)39H2O and Cu(NO3)23H2O were first dissolved in methanol. The Si/Fe molar ratio was 100/20, which corresponds to a SiO2/CuFe2O4 molar ratio of 100/10 (28.5 wt. % of CuFe2O4). The gelation time was approximately 18 hours at 45 C and the samples (pellet shape, 5 mm thick and 15 mm in diameter) were left for two days to age. Then they were dried at 40 C for three days in flowing N2-atmosphere. After drying, they were first preheated at 300 C in vacuum for two hours and then heated for four hours at various temperatures (800, 900, 1000, and 1100 C) in air. The resulting samples were then characterized using powder X-ray diffraction, Mssbauer spectroscopy, HR TEM, and magnetic measurements. Experimental techniques. A high-resolution transmission electron microscope (Topcon) was used for the direct observation of particle appearance. Particle sizes determination was carried out using Scion Images software. X-ray patterns were measured at ambient temperature using a Siemens D5000 diffractometer. Mssbauer spectra measurements were done in transmission mode with 57Co diffused into a Cr matrix as the source moving with constant acceleration. The spectrometer was calibrated by means of a standard -Fe foil and the isomer shift was expressed with re-

Preparation of CuFe2O4/SiO2 nanocomposite by the sol-gel method

699

spect to this standard at 300 K. The fitting of spectra was performed with the help of the NORMOS program. Magnetic measurements were carried out in a vibrating sample magnetometer (VSM) at 298 K.

3. Results
The samples of CuFe2O4/SiO2 nanocomposites were obtained by the sol-gel method. Previous studies indicated that vacuum treatment restrains -Fe2O3 (hematite) formation and leads to ferrite formation [7]. For this reason, the samples were first treated at temperatures up to 300 C under vacuum and then the final heat treatment at 8001100 took place. Samples annealed at 800 and 900 C were amorphous, but for those annealed at 1000 C the crystallization of the silica matrix began. This fact was confirmed by X-ray diffraction, where diffraction peaks of cristobalite appeared at 1000 and 1100 C. All of the CuFe2O4/SiO2 samples heated in the abovementioned temperature range were dark brown. 3.1. X-ray diffraction measurements and HR TEM observations All samples were characterized by X-ray diffraction measurements (step 0.1, time 50 s/step), and the results are shown in Fig. 1. X-ray diffraction patterns of samples

Fig. 1. X-ray diffraction patterns of the CuFe2O4/SiO2 samples

700

J. PLOCEK et al.

heated at 800 C indicate the presence of amorphous SiO2, manifested by the characteristic very broad diffraction at 20 (2). The recrystallization of the silica matrix into cristobalite starts at 1000 C and there is no evidence of an amorphous phase (the broad diffraction at 20 is absent) in the sample treated at 1100 C. The diffraction patterns of phases other than SiO2 exhibit broad peaks that become sharper with increasing temperature of heat treatment. This corresponds well to the crystal growth. These diffractions were found in all the studied samples annealed at the abovementioned temperatures and can be well attributed to the ferrite spinel structure. Bulk copper ferrite spinel is slightly distorted due to the Jahn-Teller (JT) effect, therefore it has a tetragonal symmetry.

Fig. 2. HRTEM of the CuFe2O4/SiO2 sample heated at 1000 C Table 1. Average particle size of the CuFe2O4/SiO2 composite depending on the annealing temperature Temperature Particle size (nm) 800 C 72 900 C 93 1000 C 153 1100 C 13017

Direct particle size observation by means of HR TEM confirms the tendency shown by the X-ray diffraction. The mean particle size of the CuFe2O4 in SiO2 nanocomposite heated at 800 C is 7 nm. Particle size rapidly increases with increasing temperature. The sample heated at 1000 C show a mean particle size of 15 nm

Preparation of CuFe2O4/SiO2 nanocomposite by the sol-gel method

701

(Fig. 2, Table 1) and the mean particle size of the spinel ferrite particles in the sample heated at 1100 C is 127 nm. Particles were very well defined and did not exhibit a diffused appearance. 3.2. Mssbauer spectra X-ray diffraction cannot distinguish between CuFe2O4 and -Fe2O3 spinel structures, especially in the case of very small particles, due to very close lattice parameters of both structures. For this reason, the measurements of Mssbauer spectra were carried out. Mssbauer spectra also yield information about the site occupation of the spinel structure and about the number of non-equivalent iron atoms.

Fig. 3. Room temperature Mssbauer spectra of CuFe2O4/SiO2 samples treated at various temperatures in the range of 8001100 C, compared to bulk CuFe2O4

Figure 3 represents spectra obtained at room temperature for samples annealed at 800, 900, 1000, and 1100 C. The sample treated at 800 C exhibits a very large
Table 2. Interpretation of the room-temperature Mssbauer spectra of the CuFe2O4/SiO2 composite heated at 1100 C Isomer shift (mm/s) 0.3610.001 0.2750.001 0.3100.002 0.3910.006 0.2250.014 Quadrupole splitting EQ (mm/s) 0.0650.004 0.0140.002 0.0430.003 0.0360.013 0.0000.825 Hyperfine field Bhf (T) 50.2960.012 47.5970.012 45.7790.022 40.8810.023 Full line width at half height (mm/s) 0.4160.008 0.4060.005 0.4550.007 0.8130.022 0.9500.033 Relative area (%) 15.60.318 40.00.899 26.70.859 14.20.395 3.40.094

Subspectrum 1 2 3 4 5

702

J. PLOCEK et al.

singlet, which is characteristic of the superparamagnetic state. In the following spectra (increasing heat treatment temperature) we can see that this band becomes broader and transforms into a sextet, which is well defined for the 1000 C annealed sample. The 1100 C heat-treated sample shows a Mssbauer spectrum that can be decomposed into four sextets and one singlet (Fig. 4). The parameters of this fit are given in Table 2. The doublet represents nanoparticles (smaller than the critical size) that are still in the superparamagetic state, but a relative area of 3.4 % suggests that almost all particles are in the ferromagnetic state. For confirming the CuFe2O4 phase in our nanocomposite, we compared the Mssbauer spectrum of pure bulk CuFe2O4 (Figure 3, top trace) with the other nanocomposites. The sextet of -Fe2O3 (hematite), which is the most stable phase at these conditions, was not found.

Fig. 4. Room temperature Mssbauer spectrum of CuFe2O4/SiO2 samples treated at 1100 C

3.3. Magnetic measurements Figure 5 shows a plot of the magnetic moment of our nanocomposites as a function of the applied field (hysteresis curves), measured at room temperature, for all the heat treatment temperatures. We can see from this figure that the saturation magnetization values increase as particle size increases with annealing temperature. The values of saturation magnetization for the prepared samples are listed in Table 3. This Table gives both the values related to the entire nanocomposites and to the ones recalculated for their pure copper ferrite components.

Preparation of CuFe2O4/SiO2 nanocomposite by the sol-gel method

703

Fig. 5. Magnetic moments of samples heated at 800, 900, 1000, and 1100 C, measured at 298 K. (Magnetic moments of the entire samples including the SiO2 matrix) Table 3. The saturation magnetic moments measured at 298 K of the CuFe2O4/SiO2 composite heated at 800, 900, 1000, and 1100 C Saturation magnetic moment (emu/g) Composite 7.9 9.2 11.2 13.4 Pure CuFe2O4 27.7 32.3 39.4 46.9

Temperature of treating (C) 800 900 1000 1100

4. Discussion
It can be seen from the powder X-ray diffraction data that the spinel phase of CuFe2O4 is formed in the silica matrix and that it is still stable in the sample annealed at 1100 C. The presence of iron (III) oxide phases was not proved. There is a theoretical possibility for the presence of -Fe2O3, because it has the same spinel structure with lattice parameters very close to those of CuFe2O4. The presence of iron oxide, however, should be accompanied by copper ferrite decomposition and the probable formation of copper silicates. On the other hand, there are no other copper compounds present in the XRD patterns, which can be considered to be indirect evidence supporting our interpretation. In addition, Mssbauer spectroscopy results clearly show that only CuFe2O4 spinel phase is present. Another finding from XRD is that the spinel phase in the composite has cubic symmetry, while pure stoichiometric CuFe2O4 is reported to be tetragonal due to the JT effect. This could be caused by the fact that the

704

J. PLOCEK et al.

diffractions of the spinel phase are broad and therefore we cannot observe the splitting of the diffraction lines due to the JT effect. Nevertheless, there are copper ferrites reported to have cubic symmetry, e.g. cuprospinel, which can be found in the mineral database of powder diffraction files [9]. It is expected that this mineral has neither the stoichiometry of pure copper ferrite nor regular occupation of the cation sites. We probably have a similar situation in our nanocomposite, which means that the structure of our copper ferrite in the silica matrix is not exactly the same same as the structure of inverse spinel. The occupation of tetrahedral and octahedral sites is rather statistical, resulting in a cubic symmetry of spinel structure. Pure CuFe2O4 is reported to have a saturation magnetization of about 25 emu/g at room temperature [10]. The ideal inverse spinel structure of (Fe)[CuFe]O4 (parenthesis means tetrahedral positions, bracket means octahedral ones) corresponds to the saturation magnetic moment of 1 B. The values of 1.32.5 B, however, have been reported in literature [11] corresponding to mixed state of spinel. These various values of and thus of saturation magnetization are supposed to be due to different cooling rates during spinel preparation. CuFe2O4 is known to have cation vacancies, whose amount varies with preparation conditions. This fact must be taken into account for the detailed interpretation of magnetic measurements. In our case, namely the study of the phase relations in the SiO2/CuFe2O4 system, we do not take these vacancies into account in the first approach to our interpretation of measurements. The calculated value of the saturation magnetization of the pure spinel ferrite phase for the 1100 C heated sample amounts to 46,9 emu/g, which is much higher than the reported value for purely inverse spinel [10]. This can be explained by the mixed character of the spinel structure; some of the copper atoms are located in tetrahedral sites. The formula of our copper ferrite can be written as: (Fe(1x)Cux)[Fe(1+x)Cu(1x)]O4. From this formula, we can write the equation for the theoretical value of as a function of the stoichiometric coefficient x. = 1 + 8x [B] From our experimental value of MS (46,9 emu/g), we can calculate the experimental value of 2,01 B per formula unit, which corresponds to x = 0.13. Therefore, the formula of our copper ferrite can be written as (Fe0,87Cu0,13)[Fe1,13Cu0,87]O4. The temperature of the Curie point for CuFe2O4 is 728 K [10], but the coercive field is very low due to a low value of magnetocrystalline anisotropy of copper ferrite. The loops are very compact. Thus, the question whether the corresponding particles in the sample are in the superparamagnetic or ferromagnetic state cannot be answered by magnetization measurements alone. Mssbauer spectra measurements must also be taken into account. From these spectra, we can see that the samples annealed at 800 and 900 C (with corresponding mean particle size of 7 and 9 nm) are superparamagnetic at room temperature, while the ones annealed at 1000 and 1100 C (with corresponding mean particle size of 15 and 130 nm, respectively) are predominantly ferrimagnetic.

Preparation of CuFe2O4/SiO2 nanocomposite by the sol-gel method

705

From HRTEM observations, the sample annealed at 1100 C contains particles with a mean size of 130 nm, which is a very high value compared to other spinel ferrites in the silica matrix that were prepared in the past [7, 8]. The characteristic values of the particle size for this annealing temperature in the case of Zn, Co, and Ni ferrites are about 1015 nm. They are probably due to higher diffusion of the Cu2+ cation in the silica matrix.

5. Conclusions
CuFe2O4/SiO2 nanocomposites were prepared by the sol-gel method and investigated using powder X-ray diffraction, HR TEM, magnetic measurements, and Mssbauer spectroscopy. The CuFe2O4 spinel phase, which is formed in the studied nanocomposite, is stable up to 1100 C. The presence of Cu2+ cations significantly affects the crystallization of the amorphous silica matrix. Crystallization starts at 1000 C and leads to a cristobalite phase without any amorphous phase at 1100 C. Magnetic measurements revealed that the spinel structure of copper ferrite has a mixed-state character. The calculated saturation magnetic moment is consistent with an approximate distribution of Cu and Fe in the tetrahedral and octahedral positions corresponding to the formula (Fe0,87Cu0,13)[Fe1,13Cu0,87]O4.
Acknowledgements This work was prepared with the financial support of a Grant from the Agency of Academy of Science, Czech Republic (KJB4032402) and Research Centrum Project of the AS CR no. LN00A028.

References
[1] TRONC E., Nuovo Cimento, 18 (1996), 163. [2] NIZNANSKY D., VIART N., REHSPRINGER J.L., J. Sol-Gel Sci. Technol, 8 (1997), 615. [3] NIZNANSKY D., REHSPRINGER J.L., DRILLON M., IEEE Trans. Magnetics, 30 (1994), 821. [4] CANNAS C., GATTESCHI D., MUSINU A., CANNAS C., GATTESCHI D., MUSINU A., PICCALUGA G., SANGREGORIO C., J. Phys Chem. B, 102 (1998), 7721. [5] DELMONTE F., MORALES M.P., LEVY D., DEL MONTE F., MORALES M.P., LEVY D., FERNANDEZ A., OCAA M., ROIG A., MOLINS E., OGRADY K., SERNA C., Langmuir, 13 (1997), 3627. [6] TRONC E., CHANAC C., JOLIVET J.P., J. Solid State Chem., 139 (1998), 93. [7] HUTLOVA A., NIZNANSKY D., PLOCEK J., BURSIK J., REHSPRINGER J.L., J. Sol-Gel Sci. Technol., 26 (2003), 473. [8] PLOCEK J., HUTLOVA A., NIZNANSKY D., BURSIK J., REHSPRINGER J.L., MICKA Z., J. Non-Cryst. Solids, 315 (2003), 70. [9] ICDD-JCPDS #25-0283 [10] SCHIEBER M.M., [in:] Experimental Magnetochemistry Nonmetallic Magnetic Materials, E.P. Wolfhart (Ed.), North-Holland Publ., Amsterdam, 1967. [11] KRUPICKA S., Fyzika feritu, Academia, Prague, 1969. Received 26 January 2005 Revised 23 March 2005